Telsys 2025: Variscite carries the value, distribution brings the volume, and the dividend squeezes flexibility

Getting to Know the Company

At first glance, Telsys looks like a veteran components distributor finally enjoying a semiconductor rebound. That is too shallow a description. In practice, this is a two-layer group: the distribution arm provides volume, customer access and local market presence, but the real economics of the group now sit in Variscite, the System-on-Module manufacturer that sells mainly abroad into end markets such as medical devices, industrial automation, agriculture and defense.

That is also the key gap between what is working and what can mislead on a first read. In 2025 consolidated revenue rose to $136.3 million and net income jumped to $36.6 million. But 90.7% of segment profit came from Variscite, while the distribution segment, despite an 84% surge in revenue, ended the year with only $4.0 million of segment profit. So the top line tells you the group is growing again. The value drivers tell you that the profit engine is still overwhelmingly Variscite.

What is working now? Variscite is growing again, its backlog rose to $46.8 million with no cancellation rights, and the company continues to benefit from reputation, technical support and working relationships with chip vendors such as NXP, Texas Instruments and Qualcomm. Distribution also recovered, largely through Nvidia-related demand, while management describes strong enterprise demand and an improving industrial trend.

What is still not clean? The active bottleneck is not demand. It is the quality of the financing behind the growth. Distribution ended the year with a $51.7 million order book, but $49.1 million of it is cancellable. Variscite, by contrast, has a cleaner backlog, yet in order to secure supply during the global memory shortage it raised inventory, expanded payables and took a new $15 million bank loan in December 2025. On top of that sits an aggressive dividend policy: Telsys paid $40 million in 2025 and declared another $10 million after the balance sheet date, while Variscite itself paid $38.8 million in 2025 and declared another $10 million after the reporting period.

That is why 2026 looks less like a breakout year and more like a two-track proof year. Distribution needs to prove that its order book converts into real sales without a wave of cancellations or another step up in customer concentration. At the same time, Variscite needs to show that it can protect supply and margins without turning inventory protection into a permanent drag on financial flexibility. The market is not signaling acute distress here: short interest stands at just 0.28% of float, far below the 1.43% level seen in November 2025.

A quick map of the business at year-end 2025:

That chart matters because it sets the right frame. Distribution was the segment that leapt higher in 2025, but Variscite remained the larger revenue contributor and by far the dominant profit engine.

One more layer of context matters before going deeper. In March 2024 Telsys completed the purchase of the remaining 28.07% of Variscite’s ordinary shares and 33.33% of its management shares in exchange for 2.286 million Telsys shares, equal at the time to 25.26% of Telsys’ share capital. So 2025 is the first full year in which the full economics of Variscite flow up to Telsys shareholders. It is also the first year in which investors can ask whether that 2024 dilution really bought accessible value rather than just a better slide deck.

Events and Triggers

The ownership story is over. The justification test has started.

The first trigger: 2025 is no longer a transitional year for Variscite inside the group. It is the first real economic test. In 2024 there was still a meaningful minority interest in Variscite, and the transaction was only completed during the first quarter. By 2025, profit attributable to non-controlling interests was down to just $141 thousand, versus $2.177 million in 2024. That means almost all of Variscite’s economics now belong at the listed company layer. The question is no longer whether Variscite is a good business. The question is whether Telsys allows that value to remain and compound rather than extracting it too quickly.

Distribution rebounded sharply, but its backlog is much less clean than the headline suggests

The second trigger: the distribution segment finished 2025 with an 84% revenue jump to $56.0 million and a total backlog of $51.7 million, versus just $13.1 million at the end of 2024. That sounds dramatic, but the quality of that backlog is far more mixed than the headline implies. Only $2.6 million is non-cancellable. The remaining $49.1 million can be cancelled without meaningful compensation. The company also says the unusual jump was driven by the fact that its main end-customer, Nvidia, placed most of the 2025 orders only in early 2025 rather than before year-end 2024.

There is a softer side to this as well. Between January 1, 2026 and February 23, 2026, cancellations on the cancellable order book amounted to about $2.0 million out of $49.2 million, or 4.1%. That was lower than the comparable post-period cancellation rate on the year-end 2024 backlog, when about $0.6 million was cancelled out of $9.3 million, or 6.6%. So the distribution backlog is still lower-quality than Variscite’s, but there is no immediate evidence of collapse.

Variscite enters 2026 with the stronger backlog, but also with rising memory costs

The third trigger: Variscite is the mirror image. Its backlog rose from $29 million to $46.8 million, and customers do not have cancellation or deferral rights. The investor presentation also describes sustained quarter-over-quarter growth since Q4 2024. That is a much stronger starting position than the one in distribution.

But there is clear operational friction underneath. From mid-2025, Variscite has been dealing with a global memory shortage. The presentation says the issue now affects all DRAM variants as well as flash components, and that quarter-over-quarter component markups can reach 40% or more in some cases. That explains why the company stopped shrinking inventory, added memory stock and leaned harder on suppliers.

Capacity does not look like the near-term problem. Procurement and working capital do.

The fourth trigger: in the presentation Variscite says it is currently using 60% to 70% of its production capacity, and that the current production site can support the addition of three lines, increasing capacity by more than 50%. In other words, the near-term bottleneck is not the factory. It sits in components, delivery security, memory pricing and inventory funding. Even the long-term expansion route is not yet resolved: Variscite was recommended for a 20,000 square meter land allocation in July 2024, but rejected the quoted valuation and is still waiting for feedback.

Efficiency, Profitability and Competition

The core story in Telsys is quality divergence. Revenue growth returned, but each of the two engines produces a very different economic profile. Distribution delivers volume, activity and customer access, yet very thin incremental profitability and high concentration. Variscite delivers the real economics, but with more procurement risk and more dependence on a small number of suppliers and channels.

Distribution: much more volume, not much more value

In numbers, distribution revenue jumped from $30.5 million to $56.0 million. Segment profit rose from $2.24 million to $4.02 million. That is a solid improvement, but it still leaves the segment on a 7.2% segment margin. At the same time, 17 customers accounted for 85% of segment sales, up from 68% a year earlier. Nvidia, through direct and indirect purchases via subcontractors, represented $35.98 million of revenue, or 26.4% of consolidated sales, versus just 11.5% in 2024.

That is the key datapoint. It explains why distribution looks much better on the growth line than on the value line. The segment benefited from a strong order wave, but it remained heavily tied to one end-customer and a contract-manufacturing chain. The revenue mix says something similar: 67% of distribution revenue came from contract manufacturers in 2025, up from 59% in 2024. This was not a broad-based, clean recovery across a wide set of end-markets. It was a sharp rebound through large industrial purchasing channels.

Variscite: less growth, much more economics

Variscite grew by a far more modest 8.2%, with revenue rising to $80.3 million. But segment profit climbed to $39.3 million and made Variscite the clear value engine of the group. Importantly, Variscite’s gross margin actually slipped to 53.1% from 55.0% in 2024, so this is not a story of margin expansion. It is a story of retaining very high profitability despite memory cost pressure.

That chart sharpens the real tension. Demand at Variscite looks stronger than it did a year ago, but every extra dollar of backlog now has to be earned under much tighter procurement discipline.

The identity of customers and suppliers matters

Here it is important to distinguish between concentration that is dangerous and concentration that is manageable. At Variscite, the Arrow group generated $27.3 million of revenue in 2025, more than 10% of consolidated revenue and about 34% of SOM segment revenue. The company says it is not dependent on Arrow because it could sell directly to customers if needed. That does not make the identity irrelevant, but it does soften the risk.

On the supplier side, NXP is deeper than a normal sourcing relationship. About 70% of Variscite’s products are based on NXP processors, and 38% of Variscite’s 2025 procurement came from NXP. That is technological dependence, not just commercial concentration, because the company explicitly says the processor cannot be changed once product development is complete. So Variscite’s reputation, early-access positioning and technical support are both a moat and a dependency.

Cash Flow, Debt and Capital Structure

This is the center of the story. If you only look at the cash balance, Telsys looks comfortable: $41.6 million of cash and cash equivalents at year-end 2025 versus $22.4 million a year earlier. But that is only part of the picture. To understand flexibility, it helps to separate two different cash frames.

Normalized cash generation: the base business still throws off strong cash

Under a normalized / maintenance cash generation lens, the picture is good. Cash flow from operations rose to $47.5 million, above net income of $36.6 million and above operating profit of $40.5 million. Cash investment in property, equipment and intangible assets was only about $2.2 million. On that basis, the underlying business still generates strong cash.

But even here it matters how that cash was built. In 2025 working capital did not consume cash on net. It supported it. Trade payables rose by $15.0 million and other payables by $9.5 million, while receivables rose by $9.2 million, other receivables by $2.5 million and inventory by $2.3 million. So operating cash flow was strong, but a meaningful part of it came from supplier financing and other short-term liabilities, not just from clean operating profitability.

The all-in picture: much less is left after the dividend

Under an all-in cash flexibility view, the picture is much tighter. Start with $47.5 million of operating cash flow, subtract $2.2 million of capitalized development and fixed investment, $1.1 million of lease principal, $3.0 million of bank debt repayment and $38.9 million of dividends to shareholders and minorities, and only about $2.4 million is left before any new borrowing.

That chart explains why higher cash is not the same as higher flexibility. The group did finish the year with more cash, but it did so while leaning harder on supplier funding and while raising a fresh $15 million bank loan in December.

The balance sheet is still strong, but no longer fully unconstrained

To Telsys’ credit, this is not a classic balance-sheet stress story. The group still had about $21.9 million of net cash at year-end, because total bank debt stood at $19.7 million against $41.6 million of cash. Equity attributable to shareholders was $74.9 million, about 51% of the balance sheet. This is not what a near-distress situation looks like.

But there is still an external signal worth paying attention to. The loan that Variscite took in June 2024 requires retained earnings not to fall below its financial liabilities at any time. The company says it is in compliance. Yet in March 2026 Variscite still needed the approval of the International Bank in order to distribute up to $10 million, and committed that retained earnings would not fall below $18 million after the distribution. That is not a distress signal, but it is also not unconstrained capital freedom.

Forecasts and What Comes Next

Four non-obvious points should frame the 2026 read:

• Telsys’ growth looks consolidated, but in reality it sits on two very different backlog types: largely cancellable distribution backlog and fully committed Variscite backlog.
• The $41.6 million cash balance looks strong, but the all-in cash picture leaves only about $2.4 million after the main cash uses of the year.
• Full ownership of Variscite made its economics more accessible to Telsys shareholders, but it also created a new question: is that value being retained for growth, or extracted too quickly through dividends?
• Capacity at Variscite does not appear to be the near-term bottleneck. Procurement, margins and working-capital discipline do.

2026 looks like a proof year, not a clean breakout year

In distribution, management expects growth in 2026 based on opening backlog and the expansion of orders from Nvidia. In the investor presentation, it also describes high enterprise demand, an industrial trend that has turned positive relative to 2024, and a large set of design opportunities in military and enterprise applications. The company is also trying to expand its foundry-services activity with GlobalFoundries.

But that should not be read as a clean breakout forecast. At the distribution level, what the market will actually measure is not just whether revenue keeps rising, but whether it rises without the backlog shrinking through cancellations and without customer concentration becoming even heavier. Put differently, this is a proof year for the quality of growth, not just the quantity of growth.

At Variscite, the test is margin and working capital

Variscite is also targeting higher sales in 2026, based on the year-end backlog and on additional orders that should come in during the year. The presentation emphasizes steady quarter-over-quarter growth since Q4 2024, the ability to meet practically all Q4 2025 delivery dates, and broader sales and vendor channels to manage the memory shortage.

But the operational condition is very clear. If higher memory costs can be passed on without a material hit to demand, and if the inventory that was bought early turns out to be a supply hedge rather than a long-term drag, Variscite can preserve both growth and profitability. If not, 2026 will show that part of the backlog was bought at a higher cash and margin cost than the headline currently implies.

There is also a supportive regulatory signal, but it does not remove the other frictions

In April 2025 the US announced a new tariff framework, and later in the year the proposed tariff rate on imports from Israel changed again. The company says that in early 2026 part of the broad tariff regime was ruled unlawful, while a new framework was then introduced. The point that matters for Telsys is that Variscite’s SOM products are currently included in an exempt annex, so as of the report date they are not subject to the retaliatory tariff. That is a supportive external signal, especially because Variscite sold $22.8 million into the US in 2025. Still, tariff relief does not solve the memory, procurement and dividend questions.

What could change the market read in the near term

Over the coming days, weeks and quarters, four issues are likely to matter most:

• The first read may be positive, because the mix of growth, $36.6 million of net income and $41.6 million of cash looks strong.
• The second read will be more selective, because $49.1 million of distribution backlog remains cancellable.
• Whether Variscite can hold margin under memory cost pressure will matter more than any general demand narrative.
• The post-balance-sheet declaration of another $10 million dividend at Telsys and another $10 million at Variscite will push the market to ask whether the company is building flexibility or maximizing extraction.

Risks

Customer concentration in distribution

The most visible risk at Telsys currently sits in distribution. Nvidia accounts for 26.4% of consolidated sales, and 17 customers make up 85% of the segment’s revenue. That does not automatically make the customer relationship negative. On the contrary, it likely reflects strong positioning inside the supply chain. But it does mean the segment looks strongest while concentration works in its favor and becomes more exposed if the ordering pattern changes.

Technological and supplier dependence at Variscite

At Variscite the risk is different. NXP is both a key supplier and a technology anchor: 70% of products are based on NXP processors, and 38% of procurement comes from NXP. That is a strength in normal times, but during a component shortage it means both product development and supply execution depend on a relatively narrow vendor base.

FX exposure without hedging

From 2025 onward Telsys reports and operates in US dollars, but part of operating expenses, including wages, rent and vehicles, remains shekel-denominated, and part of procurement is also euro-linked. As of the report date the company did not hold hedging instruments against exchange-rate exposure. So the move to a dollar functional and presentation currency simplifies the accounting picture, but does not eliminate the economic exposure.

A dividend policy that can become a flexibility drag

The last major risk is capital allocation itself. The company produced strong cash flow in 2025 and kept a healthy balance sheet, but it also paid $40 million at the parent level and declared another $10 million, while Variscite paid $38.8 million and declared another $10 million after the balance sheet date. As long as demand remains strong and the supply chain keeps working, that looks manageable. If backlog conversion slips, memory pressure deepens, or supplier funding normalizes, the same generosity could start to look like a constraint rather than a strength.

Conclusions

Telsys enters 2026 as a better company than it looked in 2024, but not as a simpler one to read. What supports the thesis is Variscite, with high profitability, cleaner backlog and a clear value engine. What prevents a cleaner thesis is the combination of a largely cancellable distribution backlog, single-customer concentration and a dividend policy that removes cash faster than it leaves room for error.

In the short to medium term, the market is unlikely to argue over whether the improvement is real. It is more likely to argue over its quality. The next reports will be read less through revenue alone and more through backlog conversion, gross margin behavior and how much financial room is still left after distributions.

Current thesis in one line: Telsys is now mostly Variscite, with a distribution arm that is growing again, but only if distribution backlog converts cleanly and dividends stop eating most of the cushion will the story look genuinely clean.

What changed versus the prior read: 2024 was a transition year after dilution and the acquisition of the remaining Variscite stake. In 2025 Variscite’s economics became fully visible at the listed-company level, but distribution returned through concentration and cancellable backlog rather than through a new clean profit engine.

The counter-thesis: the market may be too harsh on the quality question. Even if part of the distribution backlog weakens, Variscite remains a strong engine with clean backlog, high profitability and unused capacity, so the group may still keep growing without hitting real financial strain.

What could change the market read over the short to medium term: two things matter above all else: conversion of the large distribution backlog without unusual cancellations, and Variscite’s ability to protect margin and delivery timing despite memory inflation.

Why this matters: when a company effectively sells two different stories, a fast-growing distributor and an SOM manufacturer that generates almost all the profit, business quality is determined not just by how much it earns, but by who creates the value, who funds the growth and how much of that value is actually left for shareholders.

What must happen over the next 2 to 4 quarters for the thesis to strengthen, and what would weaken it: for the thesis to improve, distribution must show real conversion of its backlog without a cancellation wave, Variscite must hold a reasonable margin under memory pressure, and the company must demonstrate that dividend policy is not coming at the expense of operating flexibility. What would weaken the thesis is a sharp drop in distribution orders, another margin step-down at Variscite, or rising dependence on debt to fund distributions.